BLXBench Coding category
Code · 2026-05-10
Coding category component of the BLXBench public leaderboard. The source describes this category as implementation-focused coding tasks with structured correctness checks; scores are category score percentages.
Top models (higher is better)
| Model | Score |
|---|---|
| Claude Fable 5 | 98.5 |
| Opus 4.7 | 98.5 |
| Opus 4.8 | 98.5 |
| GPT-5.3-Codex | 98.5 |
| GPT-5.5 | 98.5 |
| Ring 2.6 1T | 96.8 |
| DeepSeek-V4-Pro | 96.2 |
| MiMo-V2.5-Pro | 96.2 |
| Qwen3.7-Max | 95.8 |
| Qwen3.7-Plus | 95.8 |
| Grok Build 0.1 | 95.4 |
| Kimi K2.6 | 95.4 |
| GLM-5.2 | 94.6 |
| Gemini 3.1 Flash-Lite | 94.3 |
| MiMo-V2.5 | 93.9 |
| DeepSeek-V4-Flash | 92.3 |
| MiniMax M3 | 92.3 |
| GLM-5.1 | 92.0 |
| Grok 4.3 | 91.2 |
| Nemotron 3 Super 120B A12B | 90.0 |
| Qwen3.6-Flash-2026-04-16 | 90.0 |
| Kimi K2.7 Code | 89.7 |
| Mistral Small 4 | 89.3 |
| Mistral Medium 3.5 | 88.9 |
| Gemini 3.5 Flash | 86.2 |
| CoBuddy | 75.5 |
| Nemotron 3 Nano 30B A3B | 73.6 |
| Granite 4.1 8B | 64.8 |
| Nemotron 3 Nano Omni 30B A3B | 62.8 |
| MiniMax M2.7 | 46.0 |
| North Mini Code | 28.0 |
| Step 3.7 Flash | 25.7 |